Good read from Anthropic Engineering lead. Advice that stuck:
"Pick unreasonable goals and tell the model that they’re very reasonable"
Makes sense as models seemingly tend to rise to the ambition of the framing you give them.
Hi, I'm Felix. I'm an eng lead working on Claude. I love watching people take flight and do things they never thought possible.
We often describe that as "being more ambitious" when using AI, which is tricky advice to follow. I wrote a tutorial:
https://t.co/vNPPs5FWm8
The $EQPT short reports are brutal. The company needs a more targeted response than what it put out. "The former employees talking bad were paid" isn't a great response to what looks like accounting fraud.
I think you can *generally* tell if a young person is genuinely curious by how they've experimented with AI. The use cases are so broad that it's bound to overlap with something they care about. Good interview question nowadays IMO.
You can literally have Claude go through earnings calls for your PA, update your model, review guidance, and flag any nuance (still getting better at that last part) while you focus on core KPIs and the overall story. Never been easier to run your own book.
Jacob is correct here—we really do earnestly believe AI could kill all humans! I personally think it is >10% within the next decade. I believe Anthropic is trying its best, but we do not yet have a plan to solve alignment for superintelligence and are not clearly on track to.
New Investment Idea: Long Smith & Wesson $SWBI: High-quality, durable business priced for 16.3% IRR through F27
- Alternative Data Model Suggests 11.5% Q3 '25 Revenue Beat (using background checks, webtraffic, and supply chain indicators)
- 1854 Lever-Action Rifleto Drive 30%+ Long Gun Revenue IncreaseThrough F27
- Improving Supply Chain Dynamics to Drive Gross Margin Uplift and Lower Inventory